Accepting the inevitable; John Ottini on the changing attitudes of developers

As the H/PC community ages, and the Pocket PC flourishes, the concept of supporting the H/PC in recent weeks seems to have started to inexorably change towards a new, community induced idea.

John Ottini contemplates the new concept, which is starting to be seen within the H/PC software developer hierarchy.

Read: Accepting the inevitable

In our language we have a saying (not a very good translation): What little Johnny used to learn, now John is using. I wanted to say that if developers decided to give their HPC software for free, users will pay back for it when they shift to e.g. Windows CE6 platform. As John (Ottini) concluded, it is bittersweet feeling but probably a smart and futureproof decision.

well, sadly, the flow of things tends that way. Maybe when billg finally figure out the right HPC OS and cajole the OEMs to release new products, the developers gotta find a way to pay their bills!

OT:
grrr ... I really wish those mozilla guys make it easier to download their source and do a build!!

I've downloaded the source for Tombo before and do a build. Slight twinkering and it build just fine, with the eVC tool. Can't the mozilla folks forego their religious fetish for commandline build/make and use the eVC tools for once!

Hi,

For less technically minded members like myself, the problem is that finding software for the correct platform is proving difficult. Secondly the software I have purchased to date doesn't work very well. Not everybody is a geek with decades of programming experience. To date I have not managed to connect my Jornada 720 to the internet.
All this doesn't encourage users to invest more in a platform. Perhaps software vendors should market their products with a bit less hype and more honesty regarding technical data and details as well as compatibility and usability.
I just don't want to spend hours in a forum like this begging others for information that should come with the software.

Maybe Linux is the way to go.

HDH

HDH

Perhaps I missed something, but how would switching to Linux solve the problems you mentioned in your forum post?

Edited by Hurricane John 2005-06-14 7:20 AM

yes, perhaps. Perhaps when linux build come with pre-fab packages and not require you to mix-match the shell(gui) pkgs with the kernel. ... AND then go and pkg_install intermediate builds (if one exist for your processor) for apps you want ... and if one does not exist for your processor and you fail to cajoule linux folks to build for you, then you have to pull down the src and build yourself.

If you do not have linux already installed on your desktop, then you may try compiling the src on your newly installed linux (on your HPC). If its already installed, then you prob won't have any problems anyway.

Of cos, there is always native linux PDAs like SHARP Zaurus! ... but it has its own quirks.

Having said all these ... I'm with you that the decision to allow HPC OEMs decide their own CPU basically mess things up. This brings us to another OT. As much as some of us dislike (Read: Abhor) the WinTel combo and wished there were more options, the HPC history showed us (in a sad way) that having multiple processor base for different OEMs can have dire consequences.

Can you imagine if right from the start, there was WinTel, MacIntel, WinMIPS, MacMIPS, WinArms, MacArms, WinSH3, MacSH3 .. .. ..

oh, did I forget ... Lintel, AmigTel, LinMIPS, AmigMIPS .. .. ..

And then after spending a bomb on a software, you bring home (or downloaded) and find out that it does not work on your processor!

Or how about, after using the latest app on your WinTel, you find the performance lack lusture, and decide to upgrade to a superior WinMIPS (), only to realise that the binary are not compatible. And the developer offers you a $10 discount for the $200 software you gotten so that you can purchase a new license for your WinMIPS machine.

My friends, we do not know our blessings with the WinTel (monopoly)

yes good point, linux won't solve your problems because it's at about pre-alpha stage and you need a lot of experience to use it. maybe later it will be easier, jlime?
how do you want to connect to the internet? dialup, lan, wlan, gsm/gprs? i can try to help you in another topic on the forum (it's offtopic here)
about software... either hpc pro ARM or hpc2000 ARM will be ok for your 720!

hahaha!
anyway i've always thought the same, that compatibility is one of the most important things in computing so monopoly is ok for me in this field...

I, speaking for myself as a hpc freak, will remember those developers that contributed in any way to our platforms (slashed prices or freeware deals) when I have to buy a *cough* retarded svga keyboardless 2ghz PPC with a 40gb hdd. Laff? I probably just spec'd WM9 devices, LOL.

Anyway, my money will go to the good willed developers who helped extend the life of my beloved hpc.I *will* do my best to avoid the developers who decide we don't exist.

OT but not. Odyessy released Pocket On-schedule 6 for the h/pc platform, keeping it fully current with their PPC version. So *they* feel the development effort was worth it. It is one sweet product. If not for $100US for a PPC/hpc or hpc/desktop bundle I'd be all over it. But I have to compare it to the cheap/freebies I can scrounge and I can't justify that cost. Besides, I'm so close to getting Agenda Fusion working on my MIPS that paying $30 for Fusion is afforable if I succeed.

I'm no programmer but if the diff between hpc2k or PPC is setting a flag at compile then a lot of developers have little excuse for dropping many hpc products. It's worse IMO when they just vault a perfectly capable program.

I'll be optomistic and hope we'll see the developers open their vaults to us.

From my exp, its really "almost" like selecting the platform to compile to. The *slight* problem is that unlike Windows desktop apps, most HPC/PPC apps are written for specific screen resolution. This explains why some PPC/HPC apps have buttons hidden by the task bar when run in HPCs.

*If* developers wrote their code to be screen-res-valent (I'm trying to say anti-ambi-valent! ), then their apps will run nicely but may still need to build two different set of binaries.

Side note: Wonder if couple of us volunteer to "mod/port" any existing (useful) PPC apps to HPC platform, *and* let the developer provide support and maintain the code, will they be willing to do that? Given that any of us involved will *sign* NDA and stuffs, *AND* they will offer the app to HPC users at an affordable (if not equal to PPC) price? What do you folks think?

Well, you're forgetting the platform API's.. if there wasn't an API difference then any CE3 app would work on any CE3 based OS.

This isn't the case 99% of the time; with unibuild applications making up a tiny percentage of applications - machine coded or core coded.

well ... for the most part (like 99% as you mentioned, though I might make that 98.746% ), apps are not using very specific platform APIs. And even if they are, in C++, a simple header file with #ifdef to "map" the APIs correctly *should* solve the problem.

Granted, there are special cases. But hey, let me throw in the gauntlet here. How about if NetFront (did I get that right?) developers accept the offer to get help from volunteers (like all the talented programmers here! ) to port their code for HPC? I think that is the sure way to tell if a particular app falls into that 1% (or 1.254%) of incompatible apps.

For reality sake, in my previous life (Read: last Job), me and a buddy started a POC for a telco back in my homeland. Their billing system was based on ARBOR (or something) running on Unix/Oracle etc. code is totally C and a teeny bit of C++. Our POC was to take a given module (chosen by them) and replace it with a .NET module and still run with all the remaining modules (on unix, oracle etc).

(Well, since this scenario is introduced here, need it be said that the POC went successfully? )

Their sys team were really like "*snigger*, go ahead". With 4000 lines of code, the C modules were ported to VB.NET in 2.5 days *and* C# in another 1.5 days. Things went smoother than their expectation (and ours too actually! ) ... and I think it kinda went a bit too smooth, 'cos I think it made their management wonder why their sys team had to spend so much time maintaining their existing code, if a port was so easy. (Did I mention that these code had to work with BizTalk as well? )

Granted, coding on eVC/eVB is not like VB.NET or C#, still I think its easier to port from PPC to HPC than from C to VB.NET and C#.

So, PPC (and ex-HPC) developers, anyone for the POC?

I think the suggestion of having the progammer types here devoting their talents to PPC conversions is a terrific idea.

But why do some PPC apps like Netfront, Wordbook, Handmap, Cash Organiser, Agenda and many others happily adapt themselves to my 800 pixel wide screen without any intervention on my part?

Is this because of better programming (read more planning) by developers or was it the lazier/easier way to write apps? These apps IMO, are top-of-their-category, regardless of platform so i can't see those developers taking the lazy way out.

I would simply love it if there was a way for developers to deposit their HPC code into escrow, or trusted third party (such as HPCFactor) in order that products may live.

Developers won't open up their source completely - if they did, so much so the better, but if people want to make money from software development, they will not open it up - then some form of trusted extension model would be useful... I do wonder if HPCFactor can work something creative out here? I'll have a chat with Chris.

The APIs underlying all WinCE3 machines are largely the same, with the UI being the real difference. The core of applications should work; it's the UI which is fundamentally different - however, writing a new UI is a non-trivial task. Couple that with the extra APIs that PocketPC developers can use (GAPI, Today screen, SVGA stuff) which is all in the UI arena, and you've got a headache in trying to make it work.

And that's before you try to make an application designed for a screen that's taller than the HPC factor fit in!

it's obvious that some developers have a better grasp on this than others - wincesoft and textmaker leap to mind (with apologies for others who I have left out). It makes me think that they have paid attention to the UI differences across platforms from day one, and write for all of 'em in a fairly straightforward manner. In the case of SoftMaker, the UI is a massive achievement IMHO... It practically replaces the standard GUI tools provided to developers!

Enough pontificating: if developers could be persuaded to deposit source into a bank which others could extend, wow... That would be brilliant. Unfortunately, realism suggests to me that it is exceedingly unlikely.

John,

With Linux you have at least access to some if not all of the source code and as end user you can have a programme modified to suit your hardware, even if it costs money. My experience with Windows applications is that once you have bought the software and opened the package, you have no longer any rights. And God forbid if you try to modify them. Have you ever bothered to read a Windows licence agreement? I am not against the protection of intellectual propert rights, far from it, but as end user I expect a piece of software to be usable.
If I purchase any other product and it does not work properly I have the right to return it to the shop, but not with software why? Such a return policy would sharpen the quality of the software and ultimately benefit the platform.

This brings me back to the my starting point. Users will only support a platform if the software works. As end user I couldn't care less what "API" and other finer details of programming is required to make it work. Nothing is more frustrating than just having bought a piece of the software and IT DOES NOT WORK.
I love HPC's and still work with my Zaurus ZR-5700 and I would dearly love to see
another hardware manufacturer bring out a new HPC.

HDH